From 79dc5d8bedd6f8d3467e919f80a61720d06b86ed Mon Sep 17 00:00:00 2001
From: Shiva Shakeri <shivashakeri@Shivas-MacBook-Pro.local>
Date: Thu, 10 Feb 2022 00:21:37 +0330
Subject: [PATCH] Initial commit

---
 .DS_Store | Bin 0 -> 6148 bytes
 KNN.py    |  67 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 67 insertions(+)
 create mode 100644 .DS_Store
 create mode 100644 KNN.py

diff --git a/.DS_Store b/.DS_Store
new file mode 100644
index 0000000000000000000000000000000000000000..c7e7bd695908a9b87657ec51363988e56a7601a5
GIT binary patch
literal 6148
zcmeHKF-`+P474Fak<z53+!sXR2TPO`2#F>iK!Gw+Xej8e;$8eKnXxUT<U&C~gT^O&
zc72}RyD84MX6DQH*E_ShnN8tD`(zj!@6$(iRT&4uvBsNyJnZ%_+ueRUs=k~c_r@OW
z3D@F`&xg8)FAias7f@CTNC7Dz1*Cu!_>}_ey|C#lQBev=0V(jU0KX3nPV9w4VthIf
z;s^koARUHroF#yb31BZA5|M#<Qh`bJI%0Uzk#AMk3x~v{o5#*rr*`%_p?K_$e2a4P
zo~S4Vq`;{H*EyZD|F7Xcoc~Wrnn?jE@UIl`$!fD&;+3kmj$V$vw!xp_toee|Fb)br
lv}0hjV@%kNZ=)#d8ppWb3x~v@BOi32eg>$EObYz90!Pew8$bX6

literal 0
HcmV?d00001

diff --git a/KNN.py b/KNN.py
new file mode 100644
index 0000000..aeba89f
--- /dev/null
+++ b/KNN.py
@@ -0,0 +1,67 @@
+from math import sqrt
+import pandas as pd
+import numpy as np
+from collections import Counter
+
+
+def trainTestSplit(data):
+    rowNum = data.shape[0]
+    splitIndex = int(80 / 100 * rowNum)
+    train = data.iloc[:splitIndex].reset_index(drop=True) #reset indexes to start from index 0
+    test = data.iloc[splitIndex:].reset_index(drop=True)
+    return train, test
+
+def eucledianDistance(p1, p2):
+    distance = np.sqrt(np.sum((p1 - p2) ** 2))
+    return distance
+
+
+
+df = pd.read_csv('wine.csv').sample(frac=1)
+
+classNum = len(set(df.iloc[:, 0]))
+trainData, testData = trainTestSplit(df)
+
+featuresTrain = trainData.iloc[:, 1:].values
+labelsTrain = trainData.iloc[:, 0].values
+
+trainNum = len(featuresTrain)
+kk = sqrt(trainNum)
+
+
+featuresTest = testData.iloc[:, 1:].values
+labelsTest = testData.iloc[:, 0].values
+
+# print(labels)
+
+def KNNPredict(featuresTrain, labelsTrain, featureTest, k=15):
+    distances = []
+
+    #for every example in the training set, calculate eucledien distance against the test example
+    for i, point in enumerate(featuresTrain):
+        distances.append((i, eucledianDistance(featureTest, point)))
+    distances.sort(key = lambda x : x[1])
+    
+    labels = []
+    for i, distance in distances[:k]:
+        labels.append(labelsTrain[i])
+
+    count = Counter(labels)
+    label = count.most_common()[0][0]
+
+    return label
+
+def confusionMatrix(predictions, labels, classNum):
+    mat = np.zeros((classNum, classNum), dtype=np.int32)
+
+    for i in range(len(predictions)):
+        mat[predictions[i] - 1, labels[i] - 1] += 1
+    
+    return mat
+
+predictions = []
+for point in featuresTest:
+    predictions.append(KNNPredict(featuresTrain, labelsTrain, point))
+accuracy = np.sum(predictions == labelsTest) / len(labelsTest) * 100
+print(accuracy)
+print(confusionMatrix(predictions, labelsTest, classNum))
\ No newline at end of file